The present invention relates to an optical disk and a method of playing back from an optical disk.
FIG. 29 is a block diagram showing a conventional optical disk recording/playback device shown in Japanese Patent Kokai Publication 114369/1992. An A/D converter 1 converts a video signal, an audio signal or the like into digital information. An information compressing means 2 serves to compress the output of the A/D converter 1. A frame sector converting means 3 converts the compressed information into sector information equal in length to a multiple of the frame period. An encoder 4 encodes the output of the frame sector converting means 3. A modulator 5 modulates the output of the encoder 4 into predefined modulated codes so as to reduce interference between codes on the recording medium. A laser driver 6 is for modulating the laser light in accordance with the modulated codes. A laser output switch 7 is driven by the laser driver 6 to vary the current supplied to the laser in an optical head 8, for emitting laser light.
An actuator 9 is for tracking the emitted light beam. A traverse or feed motor 10 is for moving the optical head 8 in the radial direction of a disk 12 which can record information by magneto-optical recording or phase-change recording.
A disk motor 11 is driven by a motor driver 19 to rotate the disk 12. The motor drivers 19 are controlled by first and second motor controllers 20. A playback amplifier 13 amplifies the playback signal from the optical head 8. A demodulator 14 demodulates the amplified playback signal to obtain data from the recorded, modulated signal. A decoder 15 decodes the demodulated signal, and a frame sector inverse conversion means 16 performs frame sector inverse conversion to restore original image data with the addresses and parities having been removed. An information expanding means 17 expands the compressed information, and a D/A converter 18 converts the expanded information into an analog video or audio signal.
FIG. 30 shows, in a simplified form, the data arrangement structure (layer structure) of the Moving Picture Coding Experts Group (MPEG) system which is a standard method of transferring and storing compressed digital moving picture information. In FIG. 30, reference 21 denotes a group of pictures (hereinafter referred to as "GOPs") consisting of information of a plurality of frames, 22 denotes a GOP layer formed of several pictures (screens), 23 denotes slices into which each picture is divided, 24 denotes a slice layer formed of several macroblocks, 25 denotes a microblock layer, and 26 denotes a block layer formed of 8.times.8 pixels.
The microblock layer 25 is a block consisting of 8.times.8 pixels, which is the minimum unit of encoding in the MPEG system, and discrete cosine transform (hereinafter referred to as "DCT") is effected taking each micro block as a unit. Four adjacent Y signal blocks and one Cb block and one Cr block which correspond, with regard to position, to the four Y signal blocks, i.e., six blocks in all form a macroblock 24. Several macroblocks 24 form a slice 23. The macroblock 24 is a minimum unit for motion compensated prediction, and the motion vector for the motion compensated prediction are determined taking each macroblock 24 as a unit.
FIG. 31 a diagram showing the conventional encoding structure for the case where 17 pictures form one GOP. In FIG. 3127 denotes an I-picture which is image information for which intra-frame DCT is effected, 29 denotes a P-picture which is image information for which forward motion-compensated DCT encoding is effected using the I-picture or another P-picture (P-picture other than the P-picture for which the forward motion-compensated DCT encoding is being effected) as a reference picture, 28 denotes a B-picture for which motion compensated DCT encoding is effected using the I-picture and/or P-pictures at preceding and succeeding positions, as reference pictures.
FIG. 32 is a diagram showing the conventional encoding structure for the case where 10 pictures form one GOP, and FIG. 33 is a diagram showing the conventional encoding structure for the case where 15 pictures form one GOP.
In the drawings, P-, B- and I-pictures are respectively represented as "P" or "P-picture," "B" or "B-picture," and "I" or "I-picture."
The operation will next be described with reference to the drawings. With the advancement in the digital image information compression technology, it is now possible to realize an image filing system which is very convenient to use, by recording the compressed information on a disk, with which search is much easier than having a VTR with a magnetic tape. Since, such disk file system handles digital information, there is no deterioration due to dubbing, and because recording and reproduction is achieved optically and there is no direct contact with the recording medium, reliability is high.
Conventionally, an optical disk recorder shown in FIG. 29 is used for recording the digital compressed motion information of the MPEG system shown in FIG. 30. The image information digitized by the A/D converter 1 is converted at the information compression means 2 into information of a standard compression picture system such as an MPEG system. The compressed information is encoded and modulated so that the effects of the interference between the codes on the disk is reduced, and is then recorded on a disk 12. By making the amount of data for each GOP substantially identical, and by dividing information into sectors having a length equal to a multiple of a frame period, editing and the like, treating each GOP as a unit, is possible.
During playback, the image information reproduced from the optical disk 12 is amplified by the playback amplifier 13, and returned into a digital data by the demodulator 14 and the decoder, and the original image data with the addresses and parities having been removed can be restored at the frame sector inverse conversion means 16. Furthermore, an image signal is restored by effecting, MPEG decoding, for example, at the information expanding means 17, and is then converted into an analog signal by the D/A converter 18 so that display on a monitor or the like is possible.
If the MPEG system is used as the digital motion compression method as described above, the encoding structure comprising one or more compressed I-pictures 27 by means of intra-frame DCT, one or more P-pictures 29 which is formed of image information obtained by DCT encoding with motion compensation in the forward direction, and one or more B-pictures 28 obtained by DCT encoding with motion compensation using I- and/or P-pictures positioned in front and at the back along the time axis, as reference pictures, as shown in FIG. 31 to FIG. 33.
Because an I-picture is obtained by intra-frame DCT, is possible to effect reproduction of the image with an I-picture independently. A P-picture, on the other hand, is obtained by forward motion compensation and the reproduction of the image with a P-picture is not effected until after the reproduction of the I-picture. Because the B-picture is obtained by prediction from both sides, both the I- and/or P-pictures must first be reproduced before the B-picture. The amount of data is the smallest and the efficiency of encoding is the best with the B-picture, because it is predicted in both directions.
Because the B-picture is not reproduced independently, it requires an I- or P-pictures, so that if the number of the B-pictures is increased, the capacity of the buffer memories must be increased, and the delay time from the data input to the image playback is lengthened. In a storage media, represented by optical disks or the like, an encoding method with a high compression efficiency is desired for long-time recording and the delay in the image playback is not problematical. Accordingly, the encoding system showing in FIG. 31 to FIG. 33 is appropriate for simple playback.
Now let us consider how the conventional image search and fast playback are effected from a disk receding data with the encoding structure as described above. If the encoding structure is as shown in FIG. 33, and if playback is made by extracting I-pictures, fast playback is possible. In this case, when an I-picture is reproduced, then a track jump is conducted to access the next or preceding GOP, and the I-picture therein is reproduced. By repeating such an operation, a fast forward or reverse playback is realized. The feed speed for fast forward or reverse playback is limited to the 15-time speed in case of FIG. 33, and 10-times in the case of FIG. 32.
In the actual image search, if the speed is too high, it is difficult for the human eyes to recognize the image. For rough recognition, the fast search at a 10-time or more speed is appropriate, but for search with regard to the details after the rough search, fast playback or reverse playback at several-time speed is necessary It is therefore necessary that special playback can be conducted over a wide range, of from several tens to several times the normal playback speed, to permit effective image search. Where the compressed data of the MPEG system is used, and if it is attempted to reproduce P-pictures in the encoding structure of FIG. 31 to FIG. 33, the B-pictures positioned before the P-pictures are also read, and it is therefore difficult to realize four to eight time speed.
Since the conventional playback method reproduce the encoding structure on-the disk as it is, special playback can be achieved only by I-pictures, and fast forward and reverse playback can be achieved only at a speed which corresponds to the number of frames contained in one GOP or a multiple thereof.
Also, with the recording format of the digital image shown in connection with the prior art examples, I-pictures, P-pictures and B-pictures are arranged in a sequence along the time axis, so that the special playback is limited to the following method.
Particularly, a fundamental method for special playback in the system for recording digital motion picture image in the prior art performs special playback using information recorded in the TOC area which is at the inner periphery of the disk. In this case, special playback is achieved by reading, in accordance with the head address of the scene change (the address of a location where a picture immediately after the scene change is recorded) or the head address of the image file recorded in the TOC area, the digital motion picture image of the I-picture stored at the address, and reproducing them in turn.
The conventional operation for reading from the optical disk in such a method is shown in the flowchart of FIG. 34. This flowchart shows the case in which special playback is effected on the basis of the address at the head of the scene in the motion picture image information recorded in the TOC area. First, a jump is made to the TOC area, and the scene head address is stored in the internal memory, and then jump is made to the address that has been stored, and the I-picture in the GOP to which jump has been made is reproduced, and displayed, and movement to the next address of jump destination is made. Such a sequence of operation is repeated.
With such a conventional method, however, a large amount of addresses which should be searched for (and to which a jump is destined) need to be stored, and the TOC information must be rewritten each time a recording is made.
Moreover, during special playback in the conventional system, it is necessary to skip B-picture data for reproducing P-pictures, but as the I-pictures, B-pictures and P-pictures are recorded on the disk in sequence, waiting time may have to be spent before reproducing a P-picture when a track jump is conducted.
Furthermore, in the conventional system the amount of data of the I-picture encoded by intra-frame DCT is larger than the amount of data of P- or B-pictures, so that super-fast playback, of several tens times the normal playback speed cannot be realized because the time for inputting data may be insufficient.
When starting a search in the conventional system for a desired GOP from an arbitrary position on the disk, the search operation must be repeated several times for finding the head of each GOP (at which time code or address of the image is recorded).
Furthermore, as the scene change position in the motion picture image information is not known, a scene-by-scene search for finding a scene cannot be achieved by using the conventional system.
In addition, because only part of the data in each GOP is read in the special playback, image playback may not be accomplished, or playback may be possible only with regard to part of the display screen in the conventional system.